1 |
On Tue, Nov 15, 2016 at 7:32 PM, Alan McKinnon <alan.mckinnon@×××××.com> wrote: |
2 |
> On 15/11/2016 21:23, Jorge Almeida wrote: |
3 |
>> On Tue, Nov 15, 2016 at 5:39 PM, Ian Zimmerman <itz@×××××××.net> wrote: |
4 |
>>> On 2016-11-14 23:52, Jorge Almeida wrote: |
5 |
>>> |
6 |
>>>> Good to know. I'm currently testing openbox without dbus-launch. No |
7 |
>>>> problem yet. |
8 |
>>> |
9 |
>>> Do you _know_ a reason you need dbus, at all? |
10 |
>> |
11 |
>> No. |
12 |
>> |
13 |
>>> |
14 |
>>> If you don't, you don't need it ;-) |
15 |
>> |
16 |
>> I would like to believe that. |
17 |
>> |
18 |
>>> |
19 |
>>> Typically, a lot of GUI apps have dbus as a soft dependency for the sole |
20 |
>>> purpose of avoiding multiple instances. So starting the app for the |
21 |
>>> second time just activates (in some general sense) the old window. |
22 |
>> |
23 |
>> Seems harmless enough. But how did they manage to convince nearly |
24 |
>> everybody that dbus is the best invention next to sliced bread? |
25 |
> |
26 |
> |
27 |
> because dbus is actually a *good* thing for gui environments more than a |
28 |
> simple window manager? |
29 |
> |
30 |
> Because ONE ipc mechanism - dbus - can replace a plethora of home-grown, |
31 |
> half-baked ipc methods that in total consume far more resources than dbus? |
32 |
> |
33 |
> dbus is a message bus, that's all it is. Simple. light, easy, gets the |
34 |
> job done in environments where lots of bits have to chat to each other. |
35 |
> |
36 |
|
37 |
Sure. Discussing details is beyond my league. Nevertheless, I cannot |
38 |
help thinking "vendor locking". Others may disagree about the |
39 |
"simple" and "light" (http://skarnet.org/software/skabus/) but I'm not |
40 |
qualified to decide... |
41 |
|
42 |
Jorge Almeida |